FUNCTION LAGS BY FUZZY SET THEORY
Hong Li · 1988
Maximum entropy spectrum estimation may fail or produce misleading results when the given autocorrelation function lags are contaminated by noise. The methods used today are complex and require the impractical prior information. In this paper, we provide a new method and a new viewpoint based upon fuzzy set theory which leads to the algorithm by solving two sets of linear inequalities which can be solved by linear programing. So the algorithm based upon the new method must converge in finite steps and must converge to the optimal solution if some regular conditions are satisfied. Simulation results show that much improvement has been obtained for the frequency resolution by the new method. Also the new method can be extended to the multidimensional case and the related fields directly.